Your CAC isn't dying in the auction. It's dying at checkout

You audit the feed. The money leaks at checkout

A paid media team in Mexico City or Guatemala City can spend three weeks pushing CPA down in Meta: new creatives, split audiences, budget shuffled between ad sets, a long argument about whether the conversion event is configured correctly. Meanwhile, between the tap on "Pay" and the order confirmation, a share of that traffic disappears — and nobody measures it with the same discipline. The loss never shows up in Ads Manager. It shows up as a higher CPA, and the easy conclusion is that the campaign isn't working.

It isn't the campaign. You're buying expensive traffic and sending it into a step that, in several markets in the region, was never designed around how people actually pay.

How Latin America pays, versus what your checkout assumes

PCMI's 2024 figures set the baseline: credit cards accounted for roughly 42% of regional ecommerce, while digital wallets and account-to-account transfers together made up about 46%. The regional average is useless for decisions, though. The country split is where copied playbooks fall apart:

• Brazil: Pix handled 40% of ecommerce in 2024 against 44% for credit cards, and is projected at 51% by 2027 — ahead of cards.
• Mexico: credit at 38% and falling. The 2027 projection is 23% credit against 46% debit.
• Chile: credit at 60%, the most card-dependent of the large markets.
• Peru: debit at 49% against 27% credit.
• Argentina: wallets at 46%, projected to 59%.

If your Mexican checkout treats credit as the default while the market shifts to debit and transfers, you don't have a creative problem — you have a payment inventory problem. In smaller markets the bias is worse, because the store usually runs on the platform's out-of-the-box configuration, built for somewhere else. Cash and bank transfer still carry real orders in Guatemala and Panama. In Venezuela, a checkout without Pago Móvil or local-currency transfer simply doesn't compete, no matter how good the ad feeding it looks.

The silent decline

The second leak is technical, and almost nobody on the marketing side looks at it. A declined payment isn't one event: there are hard declines that never recover, and soft declines that recover on a smarter retry. The most common causes in the region have nothing to do with insufficient funds:

• Cross-border acquiring. When the acquirer sits outside the issuer's country, the transaction arrives flagged as international and the local bank treats it with more suspicion. Acquiring locally lifts approval without touching a line of copy.
• Presentment currency. Charging a local card in dollars adds a layer of friction and decline risk that the customer simply reads as "my card didn't work."
• Badly implemented 3D Secure. A step-up that opens in a window the browser blocks, or that breaks inside Instagram's in-app browser, is a lost sale with the pixel already fired on InitiateCheckout.
• No retry logic. A soft decline you don't retry is money you already paid for in CPM and then handed back.

None of these four is fixable by optimizing campaigns. All four raise CAC.

Installments are a conversion lever, not a courtesy to the bank

In Mexico, Chile, Brazil and Argentina, installments — interest-free or not — aren't a payment method. They're the mechanism that decides whether a given ticket size is purchasable at all. A 4,000-peso product competes differently when it reads as "12 payments of 333," and that changes conversion on identical traffic with an identical creative. Two things go wrong almost every time: the installment price isn't shown on the product page, where the decision actually happens, and nobody checks which issuers support the promotion. If your campaign delivers mostly customers of a bank that isn't in the promo, the offer you advertised doesn't exist for them.

Your attribution can't see any of this

Here's the uncomfortable part: payment declines are invisible in the report you use to make decisions. The user arrived, fired InitiateCheckout, the platform model counted the attempt, and the order never existed. The result is an inflated CPA blamed on the channel, the audience or the creative, when the cause sits two layers below. Three things worth more than another creative test:

• Put approval rate on the same dashboard as CPM, CTR and CPA, segmented by payment method and issuing bank.
• Measure checkout conversion separately from landing page conversion. They are different problems with different owners.
• Reconcile orders against the processor, not against Ads Manager. Where the two numbers disagree, that gap is your real loss.

What to check this week

• Pull the last 90 days of approval rate, broken out by payment method and issuing bank. If nobody in the company has it on hand, you've already found the problem.
• Compare the payment methods you have enabled against the real mix in the country you're advertising in — not the mix in the case study you read.
• Run the full checkout from Instagram's and TikTok's in-app browsers, on a mid-range phone, on mobile data. That's where your paid traffic actually lands.
• Confirm whether your processor acquires locally or cross-border, and which currency the charge is presented in.
• Turn on automatic retries for soft declines and measure how much you recover.

The part nobody wants to own

Cutting CPM by 8% is a good quarter for a media team. Moving payment approval from 82% to 90% has the same effect on CAC and usually costs less — but you don't get it in Ads Manager. You get it by talking to engineering and to your processor. It's less visible work, and nobody presents it at a results meeting. It's also, in most ecommerce operations in the region, the largest untouched lever left.